In the mining industry, particularly within open-pit environments, rockfalls pose a significant threat to personnel safety, equipment integrity, and operational productivity. Geotechnical engineers are tasked with assessing the stability of rock slopes to prevent these incidents. Traditional monitoring methods, such as radar or standard visual inspections, are effective for detecting large-scale wall movements but often struggle to capture small, fast-moving rockfalls. These smaller events can still cause severe injury or damage haulage trucks, leading to costly delays. Understanding the limitations of current systems is crucial for appreciating the necessity of innovations like RockAI, which fills this gap by monitoring previously difficult-to-track areas like haul roads, waste dumps, and mine closure sites.
Government departments and national safety bodies in Australia publish extensive guidelines and accident reports regarding geotechnical hazards in mining.
RockAI utilizes thermal videography as its primary sensory input. Unlike standard cameras that rely on visible light, thermal cameras detect infrared radiation, which correlates to the temperature of objects. In a mining context, a rock breaking away from a wall often has a different thermal signature compared to the ambient air or the stable rock face behind it. This temperature difference allows the system to 'see' events even in conditions where visual visibility might be poor, such as at night or through dust. By focusing on thermal dynamics, the system provides a continuous stream of data that is distinct from visual clutter, enabling the detection of specific physical events based on heat contrast.
Australia's national science agency and geological organizations conduct research into sensor technologies and thermal imaging applications for industry.
Capturing thermal video is only half the solution; interpreting it in real-time is where RockAI innovates. The system uses Artificial Intelligence (AI) to process the video feed instantly. The AI is trained to distinguish the specific movement patterns and thermal signatures of falling rocks from other common movements in a mine, such as driving vehicles, walking personnel, or shifting machinery. This capability significantly reduces false alarms and ensures that alerts are only sent when a genuine hazard is detected. By automating this analysis, RockAI acts as a tireless observer, providing instant alerts to operators that help prevent accidents before they escalate, making the solution scalable across vast mining operations.
Universities and industry peak bodies in Australia collaborate to advance the integration of AI and robotics within the resources sector.